Windows LocalSpeech is not easy.
The tts.exe which was used always had problems.
It appended extra characters to the outputfile.
The output for an inquiry of voices was completely unstructured, and undependable.
I ended up trying a completely different way, using powershell and cortana.
It requires Windows 10 minimally, but at least the output for voices is structured.
It works on my machine - and now I have 3 voices vs previously tts.exe could only return 2.
update should be on grog-config-yaml branch ... latest build